SPWLA 54th Annual Logging Symposium, June 22-26, 2013

ESTIMATION OF POROSITY AND FLUID CONSTITUENTS FROM

NEUTRON AND DENSITY LOGS USING AN INTERACTIVE MATRIX
SCALE
Olabode Ijasan and Carlos Torres-Verdín, The University of Texas at Austin;
William E. Preeg, Consultant

Copyright 2013, held jointly by the Society of Petrophysicists and Well Log With matrix effects eliminated from the neutron-density
Analysts (SPWLA) and the submitting authors
overlay, gas- or light-oil-saturated formations,
This paper was prepared for presentation at the SPWLA 54th Annual Logging
Symposium held in New Orleans, Louisiana, June 22-26, 2013. exhibiting the characteristic gas crossover, are
emphasized to be more representative of saturating
hydrocarbons. This behavior gives a clear qualitative
ABSTRACT distinction between hydrocarbon-saturated sweet spots
and non-viable depth zones.
Neutron and density logs are important borehole
measurements for estimating reservoir capacity and
inferring saturating fluids. The neutron log, measuring INTRODUCTION
hydrogen index (HI), is commonly expressed in
apparent water-filled porosity units assuming a constant Porosity calculated from neutron and density
matrix lithology, whereby it is not always measurements is still the most reliable estimate of
representative of actual pore volume. On the other reservoir capacity from well-log analysis. In complex
hand, a lithology-independent porosity calculation from lithologies, inadequate characterization of the matrix
nuclear magnetic resonance (NMR) or/and core could yield inaccurate porosity and saturation estimates.
measurements, provides reliable evaluations of The petrophysical effects of lithology, saturating fluid,
reservoir capacity. In practice, not all wells include core and borehole conditions on nuclear logs were
or NMR measurements. The Gaymard-Poupon exhaustively discussed by Ellis and Singer (2007).
calculation of porosity from density and neutron logs, Using departure curves from log interpretation charts
especially in clean gas-bearing formations, is only an (Schlumberger, 2009), corrections are applied such that
approximation. This paper introduces an interpretation interpreted properties are representative of the
workflow wherein formation porosity and hydrocarbon formation only. Extensive studies and publications on
constituents can be estimated from density and neutron neutron and density logs, being ubiquitous for porosity
logs using an interactive, variable matrix scale and hydrocarbon estimation, can be found in the
specifically suited for the pre-calculated matrix density. literature.

Firstly, we estimate matrix components from Historically, total porosity, t, in gas-bearing
combinations of nuclear logs (photoelectric, formations is approximated with the formula (Gaymard
spontaneous gamma ray, neutron, and density) using and Poupon, 1968)
Schlumberger’s Nuclear Parameter calculator
(SNUPAR) as a matrix compositional solver while N2  D2
assuming fresh-water-filled formations. The combined t2  , (1)
2
effects of grain density, volumetric concentration of
shale, matrix-hydrogen, and neutron lithology units
where D and N are density and neutron apparent
define our interactive matrix scale for correction of
porosities, respectively. Mao (2001) studied the
neutron porosity. Under updated matrix conditions, the
correlation characteristics of D and N for distinct
resulting neutron-density crossover can only be
identification of oil- and gas-saturated zones. Spears
attributed to pore volume and saturating fluid effects.
(2006) applied lithofacies-based porosity corrections
Secondly, porosity, connate-water saturation, and
hydrocarbon density are calculated from the derived from neutron-density cross-plots for t
discrepancy between corrected neutron and density logs calculations in geologic and reservoir models. Fertl and
using SNUPAR and Archie’s water saturation equation, Timko (1971) extended the formulations of Gaymard
thereby eliminating the assumption of fresh-water and Poupon (1968) for calculation of hydrocarbon
saturation. density, hc, and detection of oil- and gas-bearing
intervals in shaly sands.

1
SPWLA 54th Annual Logging Symposium, June 22-26, 2013

The neutron-density overlay technique relies on the where e,i and Vi are electron density and volume
difference between apparent porosities, on a pre- fraction, respectively, of the i-th component up to M
defined matrix scale, for inferring hydrocarbon components. In hydrocarbon-bearing formations, t can
saturation (Shc), t, and hc. Several petrophysical be directly calculated from density logs if and only if
factors adversely affect the reliability of the overlay matrix density, m, and fluid density, f, are known
technique. For example, gas detection is challenging in precisely. Otherwise, density apparent porosity, D, is
shaly sands or shale gas reservoirs due to opposite obtained using
effects of shale-hydroxyls and gas density in overlay
characteristics. Similarly, in oil-saturated or invaded b   m
gas zones, the decreased difference between neutron D   t  D , (3)
 f  m
and density apparent porosities masks the
characteristics of light-hydrocarbon crossovers (Mao,
2001). Consequently, application of the overlay where m and f are assumed matrix and fluid densities,
technique requires implementation of a suitable matrix respectively, e.g. limestone matrix of 2.71 g/cm3 and
correction. fresh-water of 1 g/cm3. The porosity departure, D,
due to m and f assumptions (Ellis and Singer, 2007) is
In this paper, we estimate Shc, t, and hc using an both qualitatively and quantitatively intuitive such that
interactive analysis workflow based on the neutron-
b   m
    f  
density overlay technique, with explicit consideration 1
D  m , (4)
of neutron matrix scale and shale content. The analysis
  m   f  m
2 m

workflow improves reliability of the overlay technique f

in the presence of arbitrary lithology and fluid effects.

These effects and their influence on neutron and density where m and f are differences in matrix and fluid
apparent porosities are described with a synthetic densities, respectively, between assumed and true
example of known lithology and fluid constituents. The properties.
calculated hc enables the differentiation between oil
and gas-saturated intervals when hc < 0.25 g/cm3 and
hc  0.25 g/cm3, respectively, for simplified Neutron Apparent Porosity – The neutron log is an
interpretation. Our interactive matrix scale method is apparent porosity measurement, given that it refers to
applied to field examples of varying geology and an equivalent hydrogen index response in water-filled
lithology, namely carbonate and shale reservoirs, where lithology units, usually limestone. Limestone unit
porosity and fluid-saturation estimates are compared to implies an equivalent response of water-filled limestone
laboratory core measurements. formation where the pore volume equals that of the
neutron log. As shown by Gaymard and Poupon (1968),
it follows that the environmentally-corrected neutron
INTERPRETATION OF APPARENT POROSITY porosity log, N, across invaded formations can be
FROM NEUTRON AND DENSITY LOGS expressed as

M
1
The porosity value associated with neutron logs is
inherently apparent for given matrix and fluid units. On
N 
HI mf
 HI V
i 1
i i  t  N , (5)

the other hand, compensated bulk density

measurements bear no apparentness until density where HI is hydrogen index, subscript mf identifies
porosity is calculated with constant values of matrix mud filtrate, and N is porosity departure due to
and fluid properties. This is a significant difference neutron apparent porosity measurement. Equation 5
between density and neutron logs. implies that the neutron response is a superposition of
the volumetric contributions of component hydrogen
concentrations. The apparentness in N is determined
Density Apparent Porosity – The bulk density by the neutron porosity unit, usually water-filled
measurement, b, principally responds to formation limestone. For example, quartz, calcite, and dolomite
electron density such that blocks yield N of -2, 0, and 0.5 limestone pu
M
(porosity units), respectively. This matrix effect is
b   1.0704 e,i  0.188Vi , (2) qualitatively intuitive because quartz and dolomite have
i 1 lower and higher matrix densities, respectively, than
limestone. On the other hand, unlike D, the matrix
effect is quantitatively obscure and cannot be calculated

2
 SPWLA 54th Annual Logging Symposium, June 22-26, 2013

directly from equations 3 or 4. This effect is

exacerbated in complex mixtures of various lithologies.
Similarly, a gas-saturated limestone formation yields
negative N because the hydrogen index of gas is
typically lower than that of water.

Hence, a physically intuitive parameter representation

of neutron porosity responses is necessary. Using
neutron characteristic lengths, specifically migration
length, Lm, a calibration of Lm–to–neutron porosity is
used to quantify matrix and lithology response effects
(Ellis and Singer, 2007).

NEUTRON PARAMETER MODEL

The Schlumberger Nuclear Parameter calculator

(SNUPAR, McKeon et al., 1989) calculates nuclear
properties such as Lm, HI, photoelectric factor (PEF),
capture cross section, , etc., for any given mixture of
rocks and fluids. In this paper, we implement Lm for
property characterization of wireline neutron porosity
responses, typically with an AmBe (americium-
beryllium) neutron source. It follows that equation 5
can be re-written as Fig. 1. SNUPAR-calculated water-filled neutron
porosity calibration functions gs, gl, and gd for
 1  sandstone, limestone, and dolomite units, respectively.
N  gl  , (6)
The figure also shows neutron porosity responses
 Lm 
across relevant formations.
where gl represents the Lm–to–porosity calibration
function in limestone water-filled units; gl is obtained
LITHOLOGY EFFECTS
by fitting a polynomial function to the inverse of
SNUPAR-calculated Lm and limestone pore volume.
Matrix Effect – Equation 4 describes the sensitivity of
Figure 1 shows gs, gl, and gd water-filled calibration
D such that the matrix effect in water zones, i.e., when
functions for sandstone, limestone, and dolomite units,
respectively. Additionally, the SNUPAR-calculated f = 0 and f = cw (connate-water density), is given
Compensated Neutron Tool (CNT) thermal porosity by
response, shown in dashed blue, agrees well with gl.
 
Unless otherwise stated, neutron porosity logs in this   m 1
paper are expressed in limestone matrix units, where D   b    . (7)
     2  f   m  m
g l1 and g l are used to convert neutron porosity to Lm  f m 
logs, and vice versa, respectively.
In the neutron-density overlay technique, water-
Furthermore, we implement a SNUPAR-based saturated zones are expected to overlap only if the
compositional solver (Heidari et al., 2012) for matrix scale for both density and neutron corresponds
estimation of mineral and fluid concentrations from to the precise formation lithology. Otherwise, the
nuclear logs. The solver uses nonlinear minimization of neutron-density matrix effect, matrix = N - D
a constrained-error, quadratic cost function between (equations 3 and 5), depends on m and the neutron
SNUPAR-predicted properties and nuclear logs (b, N, response, N, of the matrix. Unlike D in equation 3, N
PEF) for estimation of mineral and fluid volumetric of the matrix is not quantitatively intuitive, and is only
fractions. Additionally, volumetric concentration of obtained from equation 6 by converting SNUPAR-
shale, Vsh, and water saturation, Sw, are calculated using calculated Lm to neutron porosity. Qualitatively, with a
linear scaling of the gamma ray (GR) log and Archie’s limestone matrix scale in water zones, matrix < 0
equation, respectively. across sandstone and matrix > 0 across dolomite.

3
SPWLA 54th Annual Logging Symposium, June 22-26, 2013

Shale-Hydroxyl or Matrix-Hydrogen Effect – Typically, where subscript hc identifies hydrocarbon. For gas-
shales consists of clay minerals with high hydroxyl saturated formations at reservoir conditions, one has
(OH–) content such that N > D. The shale-hydroxyl
effect, Nsh, can be approximated from equation 6 HI g  9  g 0.15  0.2  0.9   g   ,
2
(12)
using the expression  

 1  where subscript g describes gas. Equation 12 is

Nsh  Vsh gl  , (8) replicated in SNUPAR for hc = g < 0.25 g/cm3, while
 Lmsh  a SNUPAR-derived functional relationship is obtained
for oil (CnH2n+2) when hc > 0.25 g/cm3. Estimation of
where Lmsh is SNUPAR-calculated migration length in Shc, t, and hc thus requires solving equations 2, 10, 11,
shale. For example, Lmsh is approximately 15.35 cm for 12, and inclusion of a water saturation model, e.g.,
illite of density 2.78 g/cm3, whereby Nsh for pure illite Archie’s equation,
(see Figure 1), i.e., Vsh = 1, corresponds to 0.156. In
unconventional reservoirs with organic-rich kerogen aRw
matrix (Passey et al., 1990), the neutron porosity Rt  , (13)
 1  Shc 
m n
response increases due to high hydrogen content of t
organic matter. The SNUPAR-calculated HI of kerogen
could be as high as 0.8, depending on both hydrogen- where Rt is resistivity log, Rw is connate-water or mud-
carbon ratio and kerogen density. Accordingly, filtrate resistivity, a is Archie’s factor, m is porosity
equation 8 quantifies the matrix-hydrogen effect where exponent, and n is saturation exponent. It follows that
Vsh and Lmsh become Vker (volume fraction of kerogen) δ = 0 corresponds to a water or deeply invaded zone.
and Lmker (SNUPAR-calculated migration length of Consequently, the magnitudes of δ and hc dictate the
kerogen matrix), respectively. hydrocarbon type, i.e., oil or gas.

Consequently, the total matrix effect on neutron

porosity logs is an addition of matrix and Nsh, i.e., INTERACTIVE ANALYSIS OF MATRIX AND
interactive porosity departures due to apparent FLUID EFFECTS
limestone matrix scale (calculated from SNUPAR in
fresh-water-filled assumptions, equations 6, and 7) and The interactive analysis involves combined
shale-hydroxyl or matrix-hydrogen effect (equation 8). considerations of petrophysical effects due to apparent
It then follows that the corrected or re-scaled neutron matrix scale and hydrocarbon saturation. Using a
apparent porosity is given by synthetic example of a layered earth model, where well
logs are simulated with UTAPWeLS (The University of
Ncorr  N   matrix  Nsh  . (9) Texas at Austin Petrophysical and Well-Log Simulator,
Voss et al., 2009), we describe the estimation of t, hc,
and Shc using the interactive analysis workflow.
FLUID AND HYDROCARBON SATURATION
EFFECTS
Analysis Workflow – The first part of the analysis
Given equations 3, 5, 6, and 9, fluid and saturation involves matrix compositional interpretation, from b,
effects on re-scaled neutron apparent porosity can be PEF, and GR logs, using the SNUPAR-based solver
written as under the assumption of fresh-water-filled saturation.
We assume fresh-water-filled formations for two
Ncorr  t 1  Shc   , (10) reasons: (1) the environmentally-corrected N is
typically referenced on fresh-water-filled units, and (2)
to independently characterize matrix effects for
where δ is the difference in neutron response between estimation of m, given that formation fluids have
hydrocarbon-saturated and water-saturated formations. negligible or no effect on PEF and GR logs.
Several forms of equation 10 are given in Gaymard and
Poupon (1968), Mao (2001), and Quintero et al. (1998).
Using the estimated m from the matrix solver and
Gaymard and Poupon (1968) characterized δ across equation 3, we calculate density apparent porosity
invaded formations as the relative difference in HI
under the fresh-water-filled assumption, Dwf.
between residual hydrocarbon and mud-filtrate, i.e.
Accordingly, neutron apparent porosity under the fresh-
water-filled assumption, Nwf, is obtained by converting
    HI hc  HI mf  HI mf , (11) predicted Lm from the matrix solver to neutron porosity.
4
 SPWLA 54th Annual Logging Symposium, June 22-26, 2013

It follows from equations 8 and 9 that matrix + Nsh = and 2 describe the properties assumed for the synthetic
Nwf  Dwf, i.e. the interactive neutron-density lithology earth model, while Figure 3 shows the simulated
effect in limestone porosity scale, where Vsh is nuclear and resistivity logs. In Figure 4, we describe
calculated assuming linear scaling of the GR log. We the interpretation results obtained with the interactive
then calculate the corrected neutron apparent porosity, analysis workflow. Panels a, f, g, and h show that
Ncorr from equation 9 for re-scaling with D. At this estimated m, t, Sw, and hc, respectively, using the
point, the overlay characteristics of Nwf and Dwf are interactive analysis, agree well with model properties in
solely due to porosity effects, while the overlay of Ncorr Table 1. It is particularly significant that the calculated
and D is due to hydrocarbon pore volume. hc in panel h distinguishes between gas and oil-
saturated layers.
The second part of the analysis involves implementing
the SNUPAR-based solver for hydrocarbon Layers I and IV consist of water-saturated shale of illite
characterization. In this step, equations 2, 10, 11, 12, clay, whereby sh = 0.156 and shale density sh = 2.78
and 13 are solved such that a SNUPAR-defined g/cm3. After correction for shale-hydroxyl effects, the
inherent relationship between δ and hc is implemented actual matrix crossover effect, due to the shale density
in the analysis for estimation of hc, Shc, and t. The greater than limestone density, is shaded in brown in
functional relationship between HI and hc is derived panel b of Figure 4. On the other hand, layer V consists
of gas-saturated shale, such that the gas crossover effect
from SNUPAR for oil (hc > 0.25 g/cm3) and gas (hc <
becomes accentuated after correction for shale-
0.25 g/cm3). Figure 2 summarizes the interpretation
hydroxyl effect. In this layer, because gas saturation
workflow of the interactive analysis.
and Vsh impose opposite overlay characteristics, N  D
experiences a tug between gas and shale-hydroxyl
effects. This behavior in neutron-density interpretation
is especially common in logs acquired across shale gas
formations.

Layers II and III consist of gas- and oil-saturated

limestone formations, respectively. The matrix effect is
irrelevant in these layers because limestone is the
reference scale for neutron-density overlay. This is
corroborated by the overlap of Dwf and Nwf in panel c
of Figure 4. Hydrocarbons, in comparison to fresh
water, reduce the neutron porosity response because of
lower hydrogen index (equation 5). From equations 11
and 12, the hydrocarbon effect is dependent on hc and
is accentuated in gas-saturated layers when compared to
oil-saturated layers. In panels d and e of Figure 4, layer
III shows lower hydrocarbon effects and could be
inadvertently interpreted as a water-filled layer.
Consequently, the fluid solver incorporates the
resistivity measurement, Archie’s model (equation 13),
equations 2, 10, 11, and 12 for an inclusive calculation
of hc, Shc, and t. Panel h shows that the estimated hc
reliably predicts gas and oil densities in gas- and oil-
saturated layers II and III, respectively. In panel f, the t
approximation using Gaymard-Poupon’s formula
(equation 1) is valid in layer II but inaccurate in shaly
Fig. 2. Interactive analysis workflow for interpretation layers.
of neutron and density apparent porosities.
In layers VI and VII, for oil- and water-saturated
dolomite, respectively, the overlay characteristics in
panel b of Figure 4 indicate a matrix crossover. The
Synthetic Example – The interactive analysis workflow matrix effect in panel d shows that matrix = 0.0072
is described for matrix and fluid effects on density and (i.e., 0.72 pu) for sh = 0. By comparison, SNUPAR-
neutron apparent porosities, using simulated calculated CNT response yields apparent thermal
measurements across a synthetic earth model. Tables 1
5
SPWLA 54th Annual Logging Symposium, June 22-26, 2013

neutron porosity of 0.5 pu in dolomite of 0 % pore could increase due to quick spurt loss in low porosity,
volume. low pressure reservoirs (Xu et al., 2012).

Field Example II, oil-bearing shale – In this example,

FIELD EXAMPLES OF APPLICATION nuclear and array induction resistivity logs are acquired
in a well, drilled with oil-base mud, across an oil-
The interactive analysis workflow is implemented for bearing shale formation from the Eagle Ford shale play.
estimation of m, t, Sw, and hc in two field examples: Table 4 describes the assumed fluid properties and
(I) gas-bearing carbonate field of dolomite lithology Archie’s parameters for the oil-bearing shale reservoir.
where m > 2.71 g/cm3, and (II) oil-bearing shale Figure 6 shows field measurements, core
formation where m < 2.71 g/cm3. measurements, and interpreted petrophysical properties
for the oil-bearing shale example. Here, the SNUPAR-
based matrix analysis assumes kerogen (C100H100O8 of
Field Example I, gas-bearing carbonate – This field density 1.4 g/cm3), calcite, kaolinite, and illite as
example consists of conventional wireline nuclear and components of the matrix. Panel e of Figure 6
dual induction resistivity logs acquired across a gas- compares m from the interactive analysis to core
producing dolomite reservoir. Additionally, the well measurements and elemental capture spectroscopy
includes routine core measurements. Due to low (ECS) lithology analysis. The ECS-derived m (dashed
reservoir pressure, deep mud-filtrate invasion affects blue curve) is significantly larger than core m (blue
the nuclear logs and even the deep resistivity log, such circle points). This result is attributed to the exclusion
that log-derived Shc is considerably lower than in-situ of low-density kerogen matrix from the ECS analysis.
Shc for water-base mud (Xu et al., 2012). Table 3 Matrix density, m from SNUPAR-based matrix
summarizes the assumed Archie’s parameters and fluid analysis (red curve) agrees well with core
properties for the gas-bearing carbonate field. measurements. The resulting fluid crossover, in panel b
of Figure 6, after matrix-hydrogen and shale-hydroxyl
Figure 5 shows the field measurements, together with corrections, is due to combined effects of m (less than
core measurements, compared to results obtained with 2.71 g/cm3 of limestone), fluid density, and fluid
the interactive analysis. The neutron-density overlay in hydrogen index. It is found that the interactive analysis
panel b of Figure 5 emphasizes the matrix crossover yields a relatively constant hc of 0.747 g/cm3 for the
because the reservoir is primarily of dolomite lithology. interval in panel h. Furthermore, estimated t and Sw
The gas flag in panel j, proportional to hydrocarbon from the interactive analysis (panels f and g,
pore volume, is most pronounced across XY50 – XY75 respectively), agree well with core measurements.
ft despite the suppressed gas crossover in panel b.
Across the interval in Figure 5, the gas flag provides a
qualitative and unequivocal indication of hydrocarbon CONCLUSIONS
saturation despite mud-filtrate invasion and matrix
crossover. The interactive analysis workflow re-scales the neutron-
density overlay with corrected neutron and density
The calculated hc in panel h of Figure 5, with a 0.176- apparent porosities in a variable matrix scale, for
g/cm3 average, confirms that the reservoir is largely independent characterization of fluid effects. It was
saturated with gas. Conclusively, we implement found that the SNUPAR-based matrix analysis,
combined matrix and fluid volumetric analysis with the assuming fresh-water-filled formations, renders
SNUPAR-based solver, where methane gas of 0.176 accurate estimations of m even across hydrocarbon-
g/cm3 is assumed as a component of the fluids, thus saturated intervals. This result is due to the fact that
eliminating the water-filled assumption in the formation fluids have negligible or no effect on PEF
independent matrix analysis. Panel i shows cumulative and GR logs. One limitation of the SNUPAR-based
plot of the volumetric fractions of shale, quartz, calcite, matrix analysis is that a priori qualitative knowledge of
dolomite, water, and gas, obtained from the SNUPAR- matrix components, i.e., lithology, clay mineral, etc., is
based solver. The estimated m and t (panels e and f, essential for accurate estimation of m. This is achieved
respectively), agree well with core measurements. On by preliminary lithology or matrix identification cross-
the other hand, log-derived Sw (panel g) within interval plots, such as PEF-b, thorium-potassium, and PEF-
XX25 – XY00 ft is considerably lower than core potassium cross-plots (Schlumberger, 2009).
measurements. This behavior can be attributed to Furthermore, the workflow assumes minimal shoulder-
variations in Archie’s parameters for differing rock bed effects such that depth-by-depth analysis is
types along the well. Furthermore, Sw in core samples adequate for SNUPAR calculations.

6
 SPWLA 54th Annual Logging Symposium, June 22-26, 2013

The merits of the SNUPAR-based interactive analysis ILD Deep induction resistivity (-m)
workflow include the following: (1) unequivocal ILM Medium induction resistivity (-m)
identification of hydrocarbon-saturated zones, (2) Lm Neutron migration length (cm)
model-consistent formation porosity, and (3) m Archie’s porosity exponent
hydrocarbon density for gas/oil identification. It was M Number of components
shown that the workflow incorporates interactive matrix n Archie’s saturation exponent
corrections such that the Gaymard-Poupon basis and NMR Nuclear magnetic resonance
formulations for lithology-independent porosity and PEF Photoelectric factor (b/e)
hydrocarbon identification can be implemented for any Rt True resistivity (-m)
neutron-density matrix scale and lithology (clean or/and Rw Water resistivity (-m)
shaly), especially in wells with limited data. SFLU Spherically focused resistivity (-m)
SNUPAR Schlumberger Nuclear Parameter
Synthetic and field examples of application indicate calculator
that lithology-independent porosity and hydrocarbon Sw Water saturation (%)
density can be efficiently estimated from conventional UTAPWeLS The University of Texas at Austin
nuclear and resistivity logs for reliable and quantitative Petrophysical and Well-Log
detection and appraisal of hydrocarbon-saturated sweet Simulator
spots and non-viable zones. Vi Volumetric concentration (v/v)
Vsh Volumetric concentration of shale
(v/v)
ACKNOWLEDGEMENTS  Capture cross section (cu)
 Density (g/cm3)
The work reported in this paper was funded by The
 Departure
University of Texas at Austin's Research Consortium
on Formation Evaluation, jointly sponsored by Afren,  Apparent porosity (v/v)
Anadarko, Apache, Aramco, Baker-Hughes, BG, BHP m Matrix density (g/cm3)
Billiton, BP, Chevron, ConocoPhillips, COSL, ENI, t Total porosity (v/v)
ExxonMobil, Halliburton, Hess, Maersk, Marathon Oil δ Neutron fluid effect parameter ( )
Corporation, Mexican Institute for Petroleum, Nexen,
ONGC, OXY, Petrobras, Repsol, RWE, Schlumberger,
Shell, Statoil, Total, Weatherford, Wintershall, and Subscripts
Woodside Petroleum Limited. We are indebted to Shell
Oil Company for providing core and well-log b bulk
measurements used in this study. corr corrected
cw connate-water
D density
LIST OF ACRONYMNS AND SYMBOLS e electron
f fluid
a Archie’s factor g gas
AmBe Americium-beryllium hc hydrocarbon
AO10 10” array induction one foot i component index
ker kerogen
resistivity (-m)
AO30 30” array induction one foot mf mud-filtrate
N neutron
resistivity (-m)
nsh non-shale
AO90 90” array induction one foot
sh shale
resistivity (-m) t total
CNT Schlumberger compensated neutron
wf water-filled
tool
gd Neutron porosity calibration function,
in dolomite units
gl Neutron porosity calibration function,
REFERENCES
in limestone units
GR Gamma ray (API)
gs Neutron porosity calibration function, Ellis, D. V. and Singer, J. M., 2007, Well Logging for
in sandstone units Earth Scientists: Springer
HI Hydrogen index ( )
7
SPWLA 54th Annual Logging Symposium, June 22-26, 2013

Fertl, W. H., and Timko, D. J., 1971, “A distinction of

oil and gas in clean and shaly sands as derived from ABOUT THE AUTHORS
well logs,” The Log Analyst, vol. 12, no. 2.
Olabode Ijasan is a Ph.D. student in the Department of
Gaymard, R. and Poupon, A., 1968, “Response of Petroleum and Geosystems Engineering at the
neutron and formation density logs in hydrocarbon University of Texas at Austin, where received an M.Sc.
bearing formations,” The Log Analyst, vol. 9, no. 5. degree in 2010. He obtained a B.Sc. degree in Electrical
and Electronics Engineering from the University of
Heidari, Z., Torres-Verdín, C., and Preeg, W. E., 2012, Lagos, Nigeria in 2006. In 2011 and 2012, he worked
“Improved estimation of fluid and mineral volumetric as a research intern at Schlumberger-Doll Research,
concentrations from well logs in thinly bedded and Cambridge, Massachusetts, developing fast and
invaded formations,” Geophysics, 77(3), WA79-WA98. accurate methods for simulating borehole nuclear
measurements. His research interests include well-log
Mao, Z., 2001, “The physical dependence and the modeling, inversion, and petrophysical interpretation.
correlation characteristics of density and neutron logs,” He is a member of SEG, SPE, SPWLA, and IEEE.
Petrophysics, vol. 42, no. 5, 438-443.
Carlos Torres-Verdín received a Ph.D. degree in
McKeon, D. C. and Scott, H. D., 1989, “SNUPAR – A Engineering Geoscience from the University of
nuclear parameter code for nuclear geophysics California, Berkeley, in 1991. Since 1999, he has been
applications,” IEEE Transactions on Nuclear with the Department of Petroleum and Geosystems
Geoscience, vol. 36, no.1, 1215-1219. Engineering of the University of Texas at Austin, where
he currently hold the position of Zarrow Centennial
Passey, Q. R., Creaney, S., Kulla, J. B., Moretti, F. J., Professor in Petroleum Engineering, and director of the
and Stroud J. D., 1990, “A practical model for organic Joint Industry Research Consortium on Formation
richness from porosity and resistivity logs,” AAPG Evaluation. He has published more than 114 articles in
Bulletin, vol. 74, no. 12, 1777-1794. refereed technical journals, over 158 articles in
international conferences, and two book chapters. Dr.
Quintero, L. F., and Bassiouni, Z., 1998, “Porosity Torres-Verdín is co-author of two US patents. He is a
determination in gas-bearing formations,” SPE Permian member of the research committee of the SEG and was
Basin Oil and Gas Recovery Conference, March 25-27, a member of the technical committee of the SPWLA
Midland, Texas, SPE 39774. during two terms. He is recipient of the 2003, 2004,
2006, and 2007 Best Paper Award by Petrophysics,
Schlumberger Limited, 2009, Log Interpretation 2006 Best Presentation Award and 2007 Best Poster
Charts: Schlumberger. Awards by the SPWLA, SPWLA’s 2006 Distinguished
Technical Achievement Award, and SPE’s 2008
Spears, R. W., 2006, “Lithofacies-based corrections to Formation Evaluation Award.
density-neutron porosity in a high-porosity gas- and oil-
bearing turbidite sandstone reservoir, Erha field, OPL William E. Preeg received a Ph.D. degree in Nuclear
209, Deepwater Nigeria,” Petrophysics, vol. 47, no. 4, Science and Engineering from Columbia University in
294-305. 1970. From 1980 to 2002, he held various positions
with Schlumberger including Director of Research at
Voss, B., Torres-Verdín, C., Gandhi, A., Alabi, G., and Schlumberger-Doll Research (SDR), Vice-president of
Lemkecher, M., 2009, “Common stratigraphic Engineering in Houston as well as Manager of the
framework to simulate well logs and to cross-validate Nuclear Department at SDR. Prior to 1980, he worked
static and dynamic petrophysical interpretation,” for Los Alamos Scientific Laboratory, Aerojet Nuclear
SPWLA 50th Annual Logging Symposium, June 21-24, Systems Company, and the Atomic Energy
The Woodlands, Texas. Commission, largely in the area of nuclear radiation
transport. He has also served on advisory committees at
Xu, C., Heidari, Z., and Torres-Verdín, C., 2012, “Rock The University of Texas at Austin, Texas A&M,
classification in carbonate reservoirs based on static and Colorado School of Mines, and Georgia Tech.
dynamic petrophysical properties estimated from
conventional well logs,” SPE Annual Technical
Conference and Exhibition, October 8-10, San Antonio,
Texas, SPE 159991.

8
 SPWLA 54th Annual Logging Symposium, June 22-26, 2013

Layer Matrix Saturation fluid properties Interpretation comments

t = 0.10
I Illite, sh = 2.78 g/cm3 Shale and matrix effects
Sw = 1, Shc = 0
t = 0.28, Sw = 0.05, Shc = 0.95
II Limestone Gas effect
(Methane, CH4 0.182 g/cm3)
t = 0.28, Sw = 0.05, Shc = 0.95
III Limestone Hydrocarbon effects
(Oil, C16H34 0.757 g/cm3)
IV Illite t = 0.05, Sw = 1, Shc = 0 Shale and matrix effects
t = 0.10, Sw = 0.20, Shc = 0.80
V Illite Shale and gas effects
(Methane, CH4 0.182 g/cm3)
t = 0.28, Sw = 0.05, Shc = 0.95 Matrix and hydrocarbon
VI Dolomite
(Oil, C16H34 0.757 g/cm3) effects
t = 0.10
VII Dolomite Matrix effects
Sw = 1, Shc = 0
VIII Limestone t = 0 Limestone reference

Table 1. Layer properties assumed in the Synthetic Example.

Fig. 3. Simulated well logs across the synthetic multi-layer model. (a) Gamma ray (GR) log, (b) neutron and density
apparent porosities on a limestone scale, (c) array induction apparent resistivity logs, and (d) PEF log. See Table 1
for a description of assumed layer properties.
9
SPWLA 54th Annual Logging Symposium, June 22-26, 2013

Fig. 4. Interpretation results for the Synthetic Example using the interactive analysis workflow. (a) Interpreted
matrix density from SNUPAR-based matrix solver, (b) neutron-density overlay showing shale-corrected neutron log,
matrix and fluid crossover characteristics, (c) neutron and density apparent water-filled logs from SNUPAR-based
matrix solver, (d) interactive flag indicators showing matrix effect and gas flag, (e) corrected neutron-density
overlay, (f) estimated total porosity, (g) estimated water saturation, and (h) estimated hydrocarbon and fluid
densities. See Table 1 for a description of layer properties.

Variable Value Units

Connate water resistivity, Rw @ 200 °F 0.0203 -m
Connate water density, cw 1.11 g/cm3
Connate water hydrogen index, HIcw 0.936 -
Connate water salt concentration 160,000 ppm NaCl
Archie’s factor, a 1 -
Archie’s porosity exponent, m 1.95 -
Archie’s saturation exponent, n 1.75 -

Table 2. Summary of assumed Archie’s parameters and fluid properties for the Synthetic Example.

10
 SPWLA 54th Annual Logging Symposium, June 22-26, 2013

Fig. 5. Interpretation results for Field Example I, gas-bearing carbonate reservoir, using the interactive analysis
workflow. (a) Gamma ray log, (b) neutron and density porosities on limestone scale, (c) dual-induction resistivity
logs, and (d) photoelectric factor log. (e) Matrix density, (f) total porosity, and (g) water saturation from core
measurements and interactive analysis. (h) Calculated fluid densities showing a gas cut-off of 0.25g/cm3. (i)
Volumetric concentrations of rock and fluid components from SNUPAR-based solver. (j) Gas flag from interactive
analysis workflow.

Variable Value Units

Connate water resistivity, Rw @ 96 °F 0.04 -m
Connate water density, cw 1.12 g/cm3
Connate water hydrogen index, HIcw 0.932 -
Connate water salt concentration 170,000 ppm NaCl
Mud-filtrate water resistivity, Rmf @ 96 °F 0.84 -m
Mud-filtrate water density, mf 1 g/cm3
Mud-filtrate hydrogen index, HImf 1 -
Mud-filtrate water salt concentration 5147 ppm NaCl
Archie’s factor, a 1 -
Archie’s porosity exponent, m 1.96 -
Archie’s saturation exponent, n 1.83 -

Table 3. Summary of assumed fluid properties and Archie’s parameters for Field Example I, gas-bearing carbonate
(Xu et al., 2012).
11
SPWLA 54th Annual Logging Symposium, June 22-26, 2013

Fig. 6. Interpretation results for Field Example II, oil-bearing shale reservoir, using the interactive analysis
workflow. (a) Gamma ray log, (b) neutron and density porosities on limestone scale, (c) array induction resistivity
logs, and (d) photoelectric factor log. (e) Matrix density, (f) total porosity, and (g) water saturation from core
measurements and interactive analysis. (h) Calculated fluid densities showing a gas cut-off of 0.25g/cm3.

Variable Value Units

Connate water resistivity, Rw @ 215 °F 0.019 -m
Connate water density, cw 1.077 g/cm3
Connate water hydrogen index, HIcw 0.901 -
Connate water salt concentration 165,000 ppm NaCl
Archie’s factor, a 1 -
Archie’s porosity exponent, m 2.1 -
Archie’s saturation exponent, n 2 -

Table 4. Summary of assumed fluid properties and Archie’s parameters for Field Example II, oil-bearing shale.

12